FEDERAL (IN)STABILITY

Transcription

1 Chapter 6 FEDERAL (IN)STABILITY IN INDIA The conclusion appears to be that integration has not so far completely succeeded in the growth of an organic political community. In fact, politics since integration, in some cases, has tended to sharpen the sense of a separate identity. (Narain 1967, xxiii) Insufficient solidarity exists at the state level to fuel separatism. (Manor 2002, 447) It is important to point out that perceptions are more important than accurate arithmetic in influencing political behaviour. (Jeffrey 1994, 188) Federal stability is difficult to quantify. In recent years India has witnessed a proliferation of regional parties which have led commentators to express concern over the stability of the state. But federal stability should not be equated with the stability of a governing coalition. The security of a government s tenure may, of course, have implications for federal stability, but by itself the K. Adeney, Federalism and Ethnic Conflict Regulation in India and Pakistan Katharine Adeney 2007

2 108 Federalism and Ethnic Conflict Regulation longevity and stability of the government is not indicative of, nor a good proxy for, federal stability. Coalition governments do not necessarily herald federal instability; indeed, as this chapter will argue, federalism in India has become more real and stable because of the increased importance and expanding role of state-based political parties, and the necessity of building federal coalitions which reconcile regional aspirations with national cohesion (Arora 2002, 507). The success of a state in maintaining its territorial integrity is not a valid way of assessing whether a federation is stable because a state can be held together by force. In a democratic state, looking at the numbers of movements that express themselves violently and outside democratic politics is a better proxy for federal stability. Federal instability can also be measured by analyzing the number of times the center has had to deploy force to maintain normality. Religious Politics and Federal Structures Religious politics affected the operation of the federation before partition because the Muslim-majority provinces, and then the League, saw federal autonomy as a means of safeguarding their rights within a Hindu-dominated state. The demand for greater federal autonomy was rejected at the time of the Cabinet Mission Plan (CMP), leading to partition. The logic of partition was that Muslim-majority areas would be allocated to Pakistan and all other areas to India. The nominally sovereign princely states chose whether to join India or Pakistan. In practice the vast majority had no choice; their territory was surrounded by India or Pakistan. Hyderabad, Junagadh, and Kashmir were the exceptions where the ruler was a different religion than that of the majority of the population. Hyderabad and Kashmir attempted to remain independent; Junagadh acceded to Pakistan. Ultimately both Junagadh and Hyderabad were forcibly incorporated into India. Kashmir was more complicated. By virtue of its territorial position it could have acceded to either India or Pakistan, but the logic of partition would have placed it with Pakistan. In addition, the k in Pakistan stands for Kashmir and its accession was a question of national identity for Pakistan. In actuality, the Hindu Maharaja attempted to maintain the independence of the state, ultimately provoking Pashtun tribesmen allegedly supported by Pakistan to invade. This prompted the Maharaja to request Indian military assistance, which was granted on the condition of Kashmir s accession to India. This accession was conditional on a referendum being

3 Federal (In)Stability in India 109 held. The accession to India was supported by the National Conference (JKNC), which had opposed the rule of the Maharaja. Indian intervention led to the first war between India and Pakistan in 1948, which resulted in the division of Kashmir along the Line of Control. After the war, India refused to hold the promised referendum on the grounds that one third of the territory remained occupied by Pakistan. But India also had an ideological commitment to Kashmir. Despite pursuing policies that led to partition, the Congress had not accepted Pakistan s rationale for partition. For the Congress there was no logic that had to be fulfilled by the accession of Kashmir to Pakistan, which would have confirmed the Muslim League s contention that India was a Hindu Raj. Any concession permitting the breakaway of Kashmir would be tantamount to refuting the claim of India to be a secular state (Brass 1994, 192). Jawaharlal Nehru noted in 1948 that [i]f Kashmir went, the positions of the Muslims in India would become more difficult. In fact there would be a tendency of people to accept a purely communal Hindu viewpoint (Brown 2003, 213). On a personal level, Kashmir was the ancestral home of the Nehru family. 1 At the time of independence Kashmir was the only non-hindu-majority state in the Indian Union. Religious freedoms were guaranteed at the all- India level through the Constitution. Yet the situation was more complicated. As noted in Chapter 5, no self-governing rights had been granted to religious communities, but the terms of the accession to India meant that Kashmir possessed a constitutional status different from other states. Article 306A (which became Article 370), provided that the center could not legislate on any items pertaining to Kashmir other than defense, foreign affairs, and communications. This was a form of asymmetrical federalism 2 effectively guaranteeing self-governing rights to a non-hindu state. It was therefore not surprising that, after disagreements between Nehru and Sheikh Abdullah, the leader of the JKNC, over Kashmiri autonomy (which culminated in Abdullah s imprisonment) the Constitution was amended in 1954 to empower the Indian government to legislate on all matters on the union list, not just defence, foreign affairs and communication (Bose 1997, 33). Subsequent legislation enacted between 1954 and 1958 stated that Kashmir was an integral part of the Indian Union and permitted civil servants from the center to work in the state (Bose 1997, 33). Although these changes were made with the concurrence of Kashmiri Prime Minister Bakshi Ghulam Mohammed, who took over after the arrest of Abdullah, he was an unpopular prime minister and many Kashmiris resented the changes. The center was charged with breaking promises and the changes had important practical implications, as non-kashmiris became involved in the administration of the state. Sumantra Bose observes that

4 110 Federalism and Ethnic Conflict Regulation while Kashmir s political arena was monopolized by corrupt despised puppets installed at Delhi s behest its day to day administration too gradually became to be dominated by people with no roots among the population (1997, 34). These changes have been the source of grievance ever since. The perceived security implications of a weaker relationship of a contested unit with the center were paramount in the decision to undermine the asymmetrical relationship. Similar moves have been made in the Northeast. The Constitutional Order of 1954 also put drastic curbs on fundamental liberties: freedom of speech, assembly and association in the state could now be suspended at any time on grounds of security. No judicial reviews of such suspensions would be allowed (Bose 1997, 33). These decisions cannot be divorced from the Indian strategy of nation building, which perceived self-governing rights for religious communities to be problematic because of the perceived danger of secession: a legacy of partition. The undermining of self-governing rights for Kashmir did not make the insurgency of the late 1980s inevitable, but in combination with other actions of the center, it proved fatal. Any discussion of the politics of religion and federal structures cannot be divorced from the unwillingness of Nehru to sanction the creation of non- Hindu-majority provinces, which would have been a de facto concession of religious self-governing rights. The most controversial of the cases under his premiership occurred in Punjab. The States Reorganisation Commission (SRC) considered the claim carefully for a Punjabi-speaking state. They concluded that it lacks the general support of the people inhabiting the area [and] will solve neither the language problem nor the communal problem and might further exacerbate the existing feelings (1955, 146). But Nehru s considerations were very clear. He rejected the linguistic reorganization of the Punjab because it propagated communalism (Singh 2000, 90; Dua 1990, 193); [a]ny further demarcation of political boundaries within India on communal grounds was essentially non-negotiable (Brown 2003, 284). The status quo was the existence of Hindu majority provinces. As Table 6.1 reveals, with the exception of Jammu and Kashmir, no non- Hindu-majority states existed at the time of independence. Table 6.1 is a crude measure, as the existence of non-hindu-majority states says little about the politics within the borders of these states. The creation of states along religious lines is also dependent on religious communities being territorially concentrated. Since India is a Hindu-majority state, 3 it should not surprise us that the majority of its units were Hindu. Yet, when linguistic and religious identities coincided, as in the Punjab, the center was reticent about conceding territorial recognition to non-hindu-majority

5 Federal (In)Stability in India 111 areas. The religious identity became the relevant identity for the center, even if the demand was articulated around a different identity. There were other territorially concentrated areas where alternative religious groups were a majority, notably in the Christian enclaves of the Northeast. Their demands were complex, as Paul Brass acknowledges: Although the languages of the tribal peoples are entirely distinct from Assamese and although Christianity spread to many of them, language and religion were secondary issues. The main argument for separation and secession was that tribal peoples were simply not Indians at all (1994, 202). The Northeast was not reorganized in the 1950s because of this extreme diversity. As the SRC observed, Assam and north-east India seem to have been intended by nature to be the meeting place of many tribes and races (States Reorganisation Commission 1955, 183). Unlike Punjab and Kashmir, where the actions of the center were proximate in escalating the problem, the Northeast has been more complicated. This is partly because the Northeast is host to multiple claims and counter claims. As James Manor observes, its heterogeneities tend to go so far that they also undermine the politics of bargaining and with it the prospect for political stability (1998, 33; 2001, 81). However, although unique challenges existed, the conflicts have often been escalated by the actions of the center. The challenges that have been most prominent and long lasting are the demands of the predominantly Christian Nagas and Mizos. Table 6.1 Religious demographics of India s states in 1951 Group (absolute majority % Pop. Number of states % of states Difference of state s population) Hindu Muslim Christian Sikh Others No overall majority 2 n/a 2 7 n/a Total Source: Government of India (1953). Notes: 1. Jammu and Kashmir 2. Andaman and Nicobar Islands (Christians, 31 percent; Hindus, 30 percent), and Pepsu (Sikhs, 49.3 percent; Hindus, 48.8 percent)

6 112 Federalism and Ethnic Conflict Regulation This section will not repeat the history of the conflict in the region; excellent summations can be found in Bhagwan Dua (1990) and B. G. Verghese (1996). The following points are pertinent: The Nagas demanded independence at the time of partition, a demand that was rejected in no uncertain terms. At the same time, the Assamese state government violated the agreement to recognize the Naga National Council as the principal political and administrative force in the Naga Hill district and proceeded to extend its administration to the Naga area (Brass 1994, 202). The conflict escalated and in 1956 the Indian army was dispatched to suppress the secessionist rebellion, but [w]ithin a few months Nehru realised that the army in its zeal had alienated even those Nagas who were opposed to the insurgency (Dua 1990, 200). He called for the Assamese Government to grant the Nagas more autonomy. In 1960 the demand was conceded; and in 1963 Nagaland became the 16th Indian state, the one case when Nehru sanctioned the creation of a non-hindu-majority state. Unlike the demand of a Punjabi-speaking state which Nehru distrusted on the grounds that it was a proxy for the creation of a Sikh state, in Nagaland the demands were predominantly tribal and linguistic rather than religious. In the Sixth Schedule of the Constitution, tribal communities were given polyethnic rights the right to determine the language of education and self-governing rights autonomous councils. They were also given representation rights in the parliament through reserved seats, and guaranteed representation in government employment. The First Amendment to the Constitution in 1951 provided that [n]othing shall prevent the State from making any special provision for the advancement of any socially and educationally backward classes of citizens or for the Scheduled Castes and the Scheduled Tribes (Article 15). In an interesting comparison with Punjab, Nehru concluded that not conceding the demand for a Naga state threatened the integrity of India. The contrast is striking given that there had been no secessionist movement in the Punjab, but a very real one in Nagaland. A plausible explanation for Nehru s acceptance of the demand is provided by P. K. Bose, who argues that Nehru relented, in part, because the demand for a Naga state within India had come from a section of Nagas who were not advocating secession (1990, 76). Yet, although Nehru stated to his chief ministers in 1960 that [m]y conscience is at ease now, the center s hand had been forced. Nehru was not enamored with self-governing or representation rights for Scheduled Tribes (ST), opposing their retention of government employment quotas (Wilkinson 2004, 109). 4 It must be questioned whether the center would have adopted a similar attitude to a Sikh secessionist movement occurring in the 1950s. One suspects not.

7 Federal (In)Stability in India 113 In both the Naga and Mizo cases, although the challenges to the center were real, subsequent policies of the state government and the center undermined the identification with the Indian state, to say nothing of the Indian nation. The Nagas had demanded independence in 1947, but the actions of the army after the proclamation of an underground Federal Government of Nagaland (Dua 1990, 199) increased support for the demand among the wider population. 5 In the case of the Mizos, dissatisfaction existed but did not escalate into insurgency until the failure of the Assam government to provide them with timely famine relief in 1959 (Dua 1990, 203). This led to the formation of the Mizo National Front. Encouraged by the announcement of a separate state of Nagaland in 1960, they contested the 1962 elections on a platform of self-determination. After losing the election, conflict escalated, culminating with a call for secession in As in Nagaland, the Indian army was dispatched to crush the rebellion, but shortly afterward in 1972, the Union Territory of Mizoram was created. This did not prevent further conflict, Robert Hardgrave notes that [i]n the late 1970s, clashes between the Indian Army and insurgents of the Mizo National Army grew in intensity, and the Mizo National Front, supporting independence for Mizoram, was outlawed in 1979 (1983, 1174). Mizoram was finally granted statehood within the Indian Union in Although it would be misleading to consider Nehru as unmoved by electoral considerations, under Indira Gandhi the center pursued policies designed to maximize its electoral interests in the region. In 1966 the elected Naga leadership was undermined by Indira s decision to conduct talks with the secessionists (Dua 1990, 202). Most violent conflicts cannot be solved without negotiating with those holding arms. However, Dua contends that the urgency and seriousness in resolving the separatist issue appeared on Mrs. Gandhi s agenda in direct proportion to her partisan interests in the state; she could not let a non-congress state government claim credit for bringing peace to Nagaland (1990, 202). While the center has been more willing to compromise in the Northeast, partially because the overlapping nature of cleavage does not pose such a threat to the center s notion of nation building, the heterogeneity of the region has meant that these compromises have not been as successful in preventing conflict. This has meant that a quasi-martial law has been imposed in much of the Northeast reflecting both the continuing danger of unrest and the strategically vulnerable nature of the region (Hardgrave 1983, ). Even when tensions do not take a violent turn, as the Bharatiya Janata Party (BJP) recently discovered, attempts to appease the Nagas

8 114 Federalism and Ethnic Conflict Regulation through talk of a Greater Nagaland alienated its supporters in the neighboring state of Manipur (Adeney 2005, 107). Religion remains a problematic cleavage for the center to concede self-governing rights to, but as Table 6.2 demonstrates in 2006, five non-hindu states exist, and one in which no religious group has a majority. Table 6.2 Religious demographics of India s states in 2006 Group (absolute majority % Pop. Number of states % of states Difference of state s population). Hindu Muslim Christian Sikh Others No overall control 4 n/a n/a Total Source: Adapted from Government of India (1991; 1990). 1. Jammu and Kashmir 2. Meghalaya, Mizoram, Nagaland 3. Punjab 4. Arunachal Pradesh It is important to note that none of these reorganizations were conceded on grounds of religious autonomy and cannot be viewed as conceding selfgoverning rights to religious minorities. Although a contemporaneous source (Kothari 1967, 87) argued that Indira Gandhi showed a rare nerve and considerable policy initiative in reorganizing the Punjab, the redrawing of boundaries was only undertaken when the demand was more strongly couched in linguistic terms (Dua 1990, 193). The uneven treatment of communities, perceived or actual, which happened to have a non-hindu religious identity, was a demonstration of the tensions within the Indian strategy of weak religious multiculturalism. The refusal to accommodate certain groups demands was perceived to be a strategy of control and to be perpetuating historical domination. It is no accident that the majority of secessionist demands have occurred within the non-hindu periphery, the most notable being those in Kashmir, Punjab, Mizoram, Assam and Nagaland. But to attribute secessionist demands to the fact that these regions have a different dominant religion and cannot live within a Hindu India would be simplistic. These states are not particularly homogeneous along religious or linguistic lines and many of the movements in these regions have been in reaction to the central government s

9 Federal (In)Stability in India 115 inaction or manipulation (Brass 1994, ). The response of the Indian state to the demands of non-hindu religious groups has led Gurharpal Singh to argue that India is best understood as an ethnic democracy: Where non-hindu minorities have constituted a majority in the federating unit, the operation of hegemonic control has been exercised through the Hindu minority the use of residual powers by the union government; the use of administrative structures and the coercive power of the Indian state (2000, 47 48). Can the non-hindu majority of these states explain excessive central intervention? If so, we could expect the imposition of the emergency provisions of President s Rule, under Article 356 of the Constitution, to have occurred predominantly in states without a Hindu majority. The evidence is not conclusive; there is no relationship between the number of times or the number of days that President s Rule has been in force in a state and the majority religion of that state. 6 Table 6.3 demonstrates why. States, such as Punjab, Jammu and Kashmir, without a Hindu majority are at the top of the table. Significantly, however, Meghalaya and Arunachal Pradesh are at the bottom. Therefore the religious majority of the state in question does not determine the extent of the center s interference. 7 As can be seen in Chart 6.1, the imposition of President s Rule has varied dramatically between the decades. The 1970s were the decade in which the provision was used most extensively. It is no coincidence that Indira Gandhi was prime minister for most of the decade. 8 Interestingly, there is also no statistically significant relationship between the border status of a state, the length of time spent under President s Rule or even the number of times that President s Rule has been imposed. 9 Although the Indian state is obsessed with its territorial integrity, it has not exclusively intervened or suspended democratic functioning in those states that have a land border with another country. The center has been concerned with securing its control in all areas of India. This behavior has been mitigated by the Bommai Supreme Court judgment of 1994, 10 but also because India has entered an era of coalition politics, where regional parties are generally unwilling to support the use of Article 356 because future administrations may sanction its use to dismiss their own state government. The fact that there is no relationship challenges the ethnic democracy argument of Singh. Under Article 356 the center assumes the functions of the government of that state and can declare that the powers of the state legislature be exercised by the central parliament. The lack of a correlation between either of these variables with the number of days spent under

10 116 Federalism and Ethnic Conflict Regulation President s Rule is significant. These were the indicators that should have been expected to produce a significant correlation if the ethnic-democracy argument is sustainable. Although Singh s arguments concerning an ethnic democracy in India are a useful addition to the debate, he portrays the Hindu core as more homogeneous than it is (2000, Chapter 3, 45 48). As noted, it is divided along lines of region, language, and caste, which is why the BJP has been unable to secure a majority in parliament. Table 6.3 The number of days under President s Rule for the states and Union Territories of India, Name of state Days imposed No. times imposed 1 Punjab Pondicherry Jammu and Kashmir Manipur Uttar Pradesh Kerala Nagaland Gujarat Tamil Nadu Assam West Bengal Bihar Orissa Mizoram Karnataka Rajasthan Madhya Pradesh Andhra Pradesh Goa (incl. Daman and Diu) Himachel Pradesh Sikkim Haryana Tripura Meghalaya Maharashtra Arunachal Pradesh 76 1 Total Source: Data adapted from Lok Sabha Secretariat (1996), H. M. Rajashekara (1987, ), Rajya Sabha (1996; 1997; 1999a; 1999b; 2002), Dawn (2002), V. Venkatesan (2005), and Onkar Singh (2005). See Appendix 2. Notes: Non-Hindu majority states are in bold. Most lists of President s Rule do not include the Union Territories. They are included here because many of the incidents of President s Rule occurred in Arunachal Pradesh, Manipur, and Mizoram when they were Union Territories.

11 Federal (In)Stability in India 117 Chart 6.1 The number of days that President s Rule was imposed on the states and Union Territories of India by decade Source: As Table 6.3. Yet the number of days a state spent under President s Rule is not a sufficient test of the attitude of the Indian state toward religious minorities. The number of army interventions and types of violence within a state are also important in assessing where and why the Indian state has repressed secessionist movements: [t]he Indian Army has had more experience in counterinsurgency than almost any army in the world (Rajagopalan 2000, 44). Stephen Cohen has analyzed the number of occasions when the army has been called to intervene within Indian states, as Table 6.4 shows. 11 Although Cohen s data were collected before the insurgency in Jammu and Kashmir, they cover most of Indira Gandhi s tenure, during which she exploited religion for populist purposes and alienated many of the peripheral regions. These data demonstrate that 60 percent of interventions in this period occurred in Hindu-majority states. Although this challenges the argument that the Indian state has intervened militarily in its periphery more than its core, this is misleading. In 1977 only 23 percent of states and Union Territories had a non-hindu majority. The fact that they experienced 40 percent of the army interventions during this period supports the argument that religion has been important in determining the center s response to challenges. These interventions inevitably produced further alienation from the center, producing further challenges leading to military intervention.

12 118 Federalism and Ethnic Conflict Regulation Table 6.4 The causes of army intervention in India Causes of intervention Northeast, Punjab, J & K Other states Communal unrest 30% Tribal 12.5% Antiforeigner 10% Election violence 5% 2.5% Riots (uncategorized) 2.5% 2.5% Insurgency 5% Language riots 2.5% Food riots 2.5% Student riots 2.5% Rural 2.5% Caste 2.5% Other 7.5% Unknown 2.5% 7.5% Total 40% 60% Source: Data adapted from Stephen Cohen (1988, ). More recent data (shown in Table 6.5) indicates that the vast majority of deaths in major conflicts continue to occur in these regions. The left-wing violence that occurs in Hindu-majority states accounts for almost ten percent of the total, but is dwarfed by the numbers of deaths in Kashmir and the Northeast. Table 6.5 Deaths from violent conflicts in India, Civilians Security Force Personnel Insurgents Total Percentage Jammu & Kashmir Northeast Naxalite Punjab Others Total Source: Adapted from South Asia Terrorism Portal (2005a). Notes: Statistics collated until end of June This Web site reports death estimates at the lower end of the spectrum. There has been a definite relationship between the likelihood of army intervention and the border status of a state. Seventy-seven and one half percent of army interventions in the period of were within states with an international land border, as demonstrated in Table 6.6. This is significantly more than their proportion within India. This confirms the assessment that the Indian state is concerned with maintaining military control in these regions.

13 Federal (In)Stability in India 119 Table 6.6 Army interventions organized according to border status, Border status No. of Army % of army % of states in interventions interventions Indian federation No border Sea border International border Total Source: Adapted from Stephen Cohen (1988, ). The above data have demonstrated that Singh s argument is problematic. However, they do not mean that it is wrong. It is undeniable that the secessionist movements that India is most associated with Punjab and Kashmir have occurred in states that have not had a Hindu majority. However, they are also on a land border. These states have seen a very high number of deaths. In Punjab the official death-toll is around 30,000, whereas human rights groups believe that the actual fatalities are nearer 45,000 (Singh 2001, 139). In Kashmir, the figures for fatalities also vary widely. In 1996 Sumit Ganguly estimated that 20,000 insurgents, police, paramilitary personnel and civilians have lost their lives since the onset of the conflict (1996, 76). In the same year Farooq Abdullah, then chief minister of Jammu and Kashmir, put the figure closer to 50,000 (Bose 1997, 167). 12 The Northeast has also seen very high levels of conflict related to secessionist and interethnic conflicts. Between 1980 and 1986, 5,000 people were killed in Assam (Hardgrave 1993, 61; Brown 1996, 5). In 1990 L. P. Singh, the former governor of Assam, calculated that half of the deaths in India in the 1980s occurred in the Northeast of the country despite it possessing only one third of one percent of the country s total population (1990, 14). The official figures put the death toll at around 10,000 (Singh 2001, 139). The South Asia Terrorism Portal (SATPO) estimates that 14,000 people were killed in the Northeast between 1994 and 2005 (see Table 6.5). While definitive data are impossible to acquire, the numbers of deaths are much higher in these states, even when compared to other infamous incidents of violence. Official figures put the number of dead in communal riots relating to the demolition of the Babri Mosque at Ayodhya between December 6 and 13, 1992 at 1,200. Unofficial estimates at least double these figures (Jaffrelot 1996, 463). The 2002 pogrom in Gujarat killed 2,000 people. In both cases the majority of victims were Muslims. Yet even the figures at the upper range of the scale are much lower than the deaths in the states of the Northeast, Punjab, and Kashmir. The disparity is even greater when the differences in population sizes are taken into account.

14 120 Federalism and Ethnic Conflict Regulation Why have these regions seen such high levels of violence? It appears that the center s response to demands for autonomy has been conditioned by two factors: the religion of the state but also whether the state is situated next to a land border. The latter is impossible to test as there are no non-hindumajority states without a land border. Central policies toward these regions have played a large role in their exclusion from mainstream political discourse and have engendered insecurity. The army may well have intervened more in the regions of the core than the periphery, but this says nothing about the behavior of the army, nor the threats they face. Both of these are likely to be more extreme when secessionist movements are involved because the military is especially sensitive to, and scornful of, political parochialism (Cohen 1990, 197). The difference also lies in the fact that the Indian state exercise[s] more caution in repressing identities that it feels are mainstream than those which are not (Wilkinson 2002a, 23). Hardgrave agrees: [t]o say that India s national integrity is fundamentally secure is not to minimize the serious problems posed by disturbances in the Northeast and in the Punjab. These are strategically sensitive border areas: and prolonged agitation involves basic interests of national security. The government of India will do whatever it takes to bring these areas under control (1983, 1173). Ethnic difference does not cause conflict, contrary to Tatu Vanhanen s contention (1992, 14). It is the denial of recognition and accommodation that provides the conditions for conflict to flourish. The willingness of the center to intervene both politically and militarily in non-hindu-majority areas has reduced their identification with the Indian state and Indian nation. The secessionist demands are not specifically related to the heterogeneous composition of the federal units, although demands for territorial adjustment have been present in Punjab, Nagaland, and Assam. These secessionist demands have to be related to the fact that these states have been treated differently from the rest of the Union. 13 This analysis is supported by the fact that secessionist demands gained currency in the Punjab only after the Shiromani Akali Dal s (SAD) secular demands in the Anandpur Sahib Resolution of 1973 were repeatedly ignored and Indira Gandhi, seeking to divide the Sikh community to further her own electoral ambitions, supported a Sikh militant, Jarnail Singh Bhindranwale, to undermine SAD s support. Sikhs were, and are, substantially overrepresented according to their population in the Indian army a sign of integration. Bhindranwale succeeded too well, leading to a violent secessionist movement in the 1980s, which culminated in Operation Bluestar, the storming of the Sikh Golden Temple in June Sikhs were

15 Federal (In)Stability in India 121 outraged and further alienated by this action, which ultimately led to the assassination of Indira Gandhi by her Sikh bodyguards in October Atul Kohli notes that even after the anti-sikh riots in revenge for this assassination, the situation was dramatically calmed by Rajiv Gandhi, who offered broad compromises to Akalis. Elections were held in the state, Akalis came to power, and political violence came down sharply during 1985 (1997, 337). Once these concessions were withdrawn, violence again escalated demonstrating that institutional accommodation was possible. Force ultimately kept the Punjab in the Indian Union, but it was the political accommodation of the early 1990s that increased stability. Kashmir has also been treated differently from the rest of the Union. Bose quotes Nehru saying that Kashmiri politics revolved around personalities and that there was no room for democracy there (1997, 38). Despite the denial of democracy within the state 14 and its exclusion from mainstream political discourse, the state had been relatively quiescent. As Ganguly observes, we must ask not only why the insurgency occurred at all, but also why it did not occur at any earlier time, particularly during 1965 (1996, 80). Both Manor (1996, 472) and Ganguly (1996) argue that part of the explanation is related to the political awakening of the youth of the state, coupled with the decline in institutional capacity to accommodate their demands. But accommodation was still possible, as Bose argues, clearly, Kashmiris simply wanted basic democratic rights, including representative, accountable government and a voice in determining the destiny of their homeland (1997, 35). These basic democratic rights were undermined in the 1987 election. This election saw Farooq Abdullah make an alliance with the center and in the process lose many Kashmiri s support. A Muslim United Front (MUF) arose to challenge the JKNC but lost in an election widely believed to be rigged or unfair. This had a direct impact on the violent escalation of the conflict. As Bose notes, [i]t is thus not surprising that an Indian correspondent discovered after the eruption of insurgency in 1990 that nearly all the young men on the wanted list today were guarding ballot boxes for MUF (as campaign volunteers) in 1987 (1997, 46). This reading of the situation prescribes normalization of relations within these regions and a downscaling of the military presence. Despite the fact that more than 800 people lost their lives during the 2002 elections in Kashmir, the polls themselves were deemed fair and led to a change of government (BBC 2002). This was an encouraging sign. It is significant that these occurred under a Hindu-nationalist coalition government, able to

16 122 Federalism and Ethnic Conflict Regulation entertain the notion of political parties mobilizing around religious identities. The elections removed the ruling JKNC from office, a party supportive of the BJP coalition at the center. This was a striking demonstration of the changed agenda relating to this state, where elections have previously been manipulated to suit the governing party in the center and the state. The final linkage between religion and federal politics has been the rights of minorities within federal structures. One of the charges against federal structures in ethnically divided societies is that minorities within them are endangered (Nordlinger 1972, 31) because the state government is autonomous. But this argument is only valid if there are no effective consociational mechanisms at the state level or if the center is powerless to intervene to protect such minorities. If the center is merely unwilling to intervene, then federal systems of government cannot be criticized for bias against minorities; such discrimination could have occurred in any form of government. Although there is nothing inherent within a federal system of government that requires minority protection, in most federations, protection for minority communities is codified in the constitution. India is no exception. However, as Brass (1982, 228) points out, the maintenance of minority rights depends on the cooperation of the state governments. Steven Wilkinson (2004) has stressed this point more recently, especially in relation to law and order: all states are constitutionally required to permit religious communities to set up their own educational institutions. The Constitution also prohibits discrimination in the allocation of public funds to religiousminority schools. But, as Wilkinson eloquently discusses (2004, 102), these provisions have often been ignored. This was memorably seen during the anti-sikh riots in Delhi following the assassination of Indira Gandhi in 1984, the destruction of the Babri Masjid Mosque in Ayodhya in 1992, and the pogrom against Muslims in Gujarat in early In the latter case, Chief Minister Narendra Modi s BJP administration was widely condemned for, at best, failing to prevent the violence and, at worse, aiding and abetting it (Human Rights Watch 2002). During the violence, up to 2,000 people, mainly Muslims, died (Ali Engineer 2003). Although the above examples are extreme ones, many other examples of discrimination against minority communities exist. Judith Brown notes that the fate of minorities rested not on [Nehru] as visionary, nor even on the provisions of the constitution, but on the attitudes and practices of state governments (2003, ) and observes his frustration with this situation. Wilkinson records that [w]ithin three years of independence most Indian states abolished rules that guaranteed Muslim, Sikh and Christian proportionality in politics and employment (2004, 102, 109) and that

17 Federal (In)Stability in India 123 Uttar Pradesh (UP) refused to adopt Urdu as a state language despite the large community of Urdu speakers within the state. The Bihar government acted in a similar fashion. The center also devolved controversial issues to the states for example, policies relating to cow slaughter. Nehru contended that the 1955 Bill introduced in the Lok Sabha fell under the jurisdiction of state legislatures (Mitra 1991, 771). Although this was a strategy to head off an all-india law, it did nothing to remove the issue from political discourse and merely devolved power to the level of government more likely to ban the practice as several state governments subsequently did (Wilkinson 2004, 117). The issue was an important one for the Muslim community; an all-india law would have confirmed that the Indian state was a Hindu Raj. But the actions of the state governments did little to dispel this fear. 16 In this context, the behavior of the All India Anna Dravida Munnetra Kazagham (AIADMK) government, in Tamil Nadu in 2002, appears less of an aberration. Jayalalitha Jayaram s decision to introduce an anticonversion bill banning religious conversions by allurements or force in Tamil Nadu (Krishnakumar 2002) 17 was an attempt to attract the BJP back into an electoral alliance as well as capitalizing on the BJP s vote bank. But antiminority policies at the state level have a long history. In relation to this particular example, similar laws were passed in Orissa (1967), Madhya Pradesh (1968), and Arunachal Pradesh (1977). These laws were challenged for their constitutionality. Denying the right of propagation negates Article 25 of the Constitution. This provides that all persons are equally entitled to freedom of conscience and the right freely to profess, practise and propagate religion subject to public order, morality and health. But the Supreme Court in 1977 observed that [w]e have no doubt that what the Article grants is not the right to convert another person to one s own religion, but to transmit or spread one s religion by an exposition of its tenets (Aruna 2002). The court thereby upheld the constitutionality of the laws by making a distinction between propagation of religion and conversion to that religion (Thampu 2002). 18 The above discussion demonstrates that federal politics have not always accommodated religious minorities in the same manner as they have linguistic ones. Many state governments have acted against their religious minorities or refused to act to protect them. And the center has also failed to act. Wilkinson persuasively argues that the likelihood of violence is the result of police action or inaction, dictated by the commands of their political masters the state governments. He argues that in situations where the state government requires the support of the minority community, they have an incentive to prevent violence from occurring and to take minority cultural

18 124 Federalism and Ethnic Conflict Regulation rights more seriously. This was why Urdu was reestablished as a state language in Bihar and UP in the early 1980s (2004, 126). Thus, policies at both the center and state level have been discriminatory or failed to act neutrally. Narendra Subramanian observes that the center has used more repression against Kashmiri nationalists ambivalent about being part of India than against Hindu revivalists, although the former (unlike the latter) rejected a politics of religious identity, built coalitions across religious lines, and did not incite systematic violence against outgroups until the late 1980s. This was because Hindu revivalists, unlike Kashmiri nationalists, were determined foes of secession (1999, 721). Similar reasoning explained the policies toward Punjab. This contrasts drastically with the policies adopted to protect and institutionalize language. Linguistic Politics and Federal Structures Nehru accepted minority languages as a means of provincial communication, but he sought to marginalize linguistic identities within decision-making institutions. He ultimately conceded linguistic reorganization, primarily for electoral reasons. Linguistic reorganization in India was a necessary concession to preserve the Congress s electoral hegemony. On another level, linguistic reorganization can be understood as the element of federal design that has increased India s federal stability. However, India has not been free from linguistic conflicts. The tensions relating to the choice of national language were ultimately resolved, but other conflicts persisted. The nonrecognition of Urdu in Bihar and UP has already been discussed. Wilkinson notes that in autumn 1947 the provincial UP and Central Provinces governments decided that Hindi in the Nagari script would henceforth be the only acceptable language for government business. Bihar followed suit (2004, 116). States set their own languages and were encouraged to include as state languages those languages that were spoken by a substantial proportion of the states population. 19 In the case noted above, the lack of recognition of Urdu was a religious slight. The Constitution provides for linguistic protection for minorities under Articles 29 and 30. In 1949 the Provincial Education Ministers conference determined that the State Governments are required to provide instruction of children in the primary stage in their mother tongue provided that there are at least 40 pupils speaking that language in the whole school or 10 in one class. Facilities

19 Federal (In)Stability in India 125 must also be provided in the secondary stage, if there is a sufficient number, usually one third of the total number of pupils (States Reorganisation Commission 1955, ). These provisions were approved. But although Jyotirindra Das Gupta has claimed that language demands of these minorities are usually directed to securing facilities of instruction in the mother tongue, and that [s]ince they claim a small part of the respective state s resources, it has not been difficult to reach a negotiated settlement (1975, 486), tensions have existed. Das Gupta argues that federal intervention produced the desired result in the case of Urdu in UP (1975, 486). However, a different reading of the situation reveals the failure of central intervention to persuade the state assemblies to adopt Urdu as a state language and the adoption of Urdu only when the votes of minorities became important to the state government (Wilkinson 2004, 126). As Brass notes, the major linguistic and ethnic problems of India today concern the status of minority languages, religions, and ethnic groups within the linguistically reorganized states (1982, 228). He argues that the central government has favored pluralism, but has been unable to impose its wishes on recalcitrant states. In 1980, 16 percent of Bengali speakers, 11 percent of Kannada speakers, 24 percent of Punjabi speakers, and 17 percent of Telugu speakers lived outside of their linguistic state (Kamat 1980, 1054). These particular linguistic communities have been subject to discrimination ranging from the nonprovision of educational institutions and the refusal to provide government exam papers in their mother tongue, to not even printing the rights of linguistic minorities within a state in the language of that minority. Wilkinson notes that the Linguistic Minorities Commission in the mid 1960s uncovered discrimination against Bengali, Urdu and Oriya speakers in Bihar, Telugu and Kannada speakers in Tamil Nadu, Punjabi speakers in Haryana, and Hindi speakers in Tamil Nadu (2004, 116), and Brass noted a similar trend a decade later (1982, 228). The similarity to the groups listed by A. R. Kamat is striking, but smaller communities have also suffered. The case of Assam deserves special note. In the 1951 census Assamese speakers comprised 59 percent of the population of the state, and Bengali speakers comprised 19 percent. The first conflict in the state was a linguistic one: a need to define the state as Assamese and to claim preferential polices in jobs (Baruah 1986, ), as Bengalis have long dominated [the] Assam state administration (Hardgrave 1983, 1175). Assamese was adopted as the official language of the state despite the concerns of the tribal population and the Bengali Hindus. Interestingly, given later developments [t]he Bengali Muslim immigrants allied with the ethnic Assamese

20 126 Federalism and Ethnic Conflict Regulation on cultural policy issues, while Bengali Hindus were among the most vociferous opponents (Baruah 1986, 1191). The Sons of the Soil movement that occurred in 1979 was a reaction to the increased immigration into Assam, partially because of Bangladesh s secession from Pakistan. Many Assamese were concerned about becoming a minority in their own state. Assam s population rose by 505 percent between 1901 and 1981 compared to 187 percent in the whole of India (Singh 1987, 265); much of this increase was the result of immigration. The movement demanded the deportation of primarily Bengali-speaking foreigners, defined as anyone who had entered the state after The issue had religious implications. To expel Hindu immigrants would have alienated significant sections of Hindu opinion [but] to explicitly distinguish between Hindu refugees and Muslim illegal aliens would have cut into the secular fabric of the state and would have alienated India s Muslim minority (Baruah 1986, 1192). The conflict took on an increasingly communal character with the involvement of Hindu and Muslim organizations (Hardgrave 1983, 1174), and the issue remains essentially unresolved today, despite the 1985 accord (Sen 2003; Adeney and Lall 2005, 277). Despite the tensions noted above, federal recognition of language promoted security for regionally dominant linguistic communities and restructured identity politics. As discussed in Chapter 1, the creation of homogeneous states within a federation accentuates other divisions: once a self is secure, it becomes more difficult to determine who the self is (Horowitz 1985, 617). Individuals are comprised of multiple, overlapping, and sometimes complementary identities. Once one identity is secure it provides the conditions for other identities to come to the fore. Under such conditions in a democracy, parties proliferate reflecting or mobilizing alternative identities. In these cases, the area of political competition primarily shifts within the state rather than between the state and the center. 20 This is shown markedly in the number and type of parties in the party system. Political parties are relatively easy to count, although measures to assess their relative weight are contested (Sáez 2002, 47 49). Party systems within federations have a regional and a national dimension. This study concentrates on the national dimension. 21 Much of the political-science research on India has concentrated on political parties, including Myron Weiner (1957), Iqbal Narain (1967), Stanley Kochanek (1968), and Rajni Kothari (1970). Attention has only increased with the multiplication of parties, first at the state level and then at the center. In recent years many authors have posited a relationship between the party system and the federal system in India (Vanhanen 1992; Manor 1995; Verney 1997; Chhibber and Kollman 1998; Rudolph and Rudolph 2001;

21 Federal (In)Stability in India 127 Wyatt 2001; Arora 2002; Sridharan 2002). It is not hard to fathom the reason for this. Twenty years ago, William Riker discussed India s nonconformity with Maurice Duverger s Law, 22 partly because of its one party system at the center, but also its multiple parties at state level, despite using the simple plurality electoral system (1982, 761). Simple plurality worked to the benefit of the Congress, the party with the widest and largest organizational capacity after independence. It magnified Congress votes into a majority of seats in the elections of and , despite the fact that the Congress never won a majority of votes at the all-india level. In the 1990s the party was no longer the beneficiary of the system. This was demonstrated most dramatically in 1999 when the party gained 28.3 percent of the votes 4.6 percent more than the BJP but secured 68 fewer seats. 23 Table 6.7 Number of recognized parties in Indian general elections, Election Year National Parties State Parties Total (not including Effective number independents) of legislative parties Source: Data on the numbers of national parties adapted from Election Commission of India (ECI) ( ). Data on the effective number of legislative parties adapted from David Butler (1995) and ECI ( ). As Table 6.7 illustrates, an extraordinary number of parties competed in the first election of Several authors have argued that the large number of parties who opposed the Congress did not have an incentive to merge because the Congress was in such a commanding position (Riker 1982, 761; Cox 1997, 77). Although this explanation makes intuitive sense, it does not account for the huge drop in the numbers of parties between the elections of 1952 and The electoral system remained constant, and therefore the drop in parties is best explained by the fact that the majority of these parties overestimated their support in 1952 (Weiner 1957, 227).

22 128 Federalism and Ethnic Conflict Regulation The electoral system has undoubtedly played a large role in influencing the number of parties, initially reducing their success rate. In 1952 there were 53 parties, and in 1957, after the consolidation of the party system, only 15. After this initial drop, the number of parties competing in elections has dramatically increased over time. In 1999 there were 230 parties plus many independent candidates. In 2004 the number of parties has increased exponentially. The effective number of legislative parties has also risen dramatically from 1.78 in 1952 to 6.35 in The creation of a multiparty system in a simple plurality electoral system is problematic for institutionalist theorists of party systems. The percentage of seats gained by state parties at the center increased from 16 percent in 1991 (Rudolph and Rudolph 2001, 1543) to 27 percent in Coalition governments now look to be the norm rather than the exception. The 2004 elections reinforced the continuing importance of pre-poll bargaining, vote pooling in Donald Horowitz s terminology (1985, 386), rather than postpoll seat pooling, although this remains important. As will be discussed in Chapter 7, the effective numbers of legislative parties have also increased over time in Pakistan. Both countries retained the simple plurality electoral system but changed their federal structures through linguistic reorganization and the One Unit Plan. The different types of provincial design affected federal stability; the effects of which can be seen through the development of the party system: the number and types of parties. As well as being an exception to institutionalist theories of party systems, India has also been an exception to sociological theories. Sociological theorists argue that party systems reflect the primary cleavages in a society such as class, language, or religion. The most famous exponents of this argument are Seymour Martin Lipset and Stein Rokkan who concentrate on conflicts and their translations into party systems (1990, 93). India was historically an exception. There are many divisions within Indian society, yet the Congress attracted a large section of the electorate across many diverse groups (Chhibber and Petrocik 2002, 56). Sociological theories appear to be more relevant to India from the 1990s with caste and regional politics coming to the fore, yet there is no primary cleavage in Indian politics. Those who view federal structures as destabilizing if they give autonomy to homogeneous units often do so because they predict that the identity around which that unit was created will become the primary cleavage between that unit and the center, or between that unit and other units. This cleavage is said to enhance instability, at best, and encourage separatism or secessionism, at worst. But in India, strong linguistic parties have not emerged in the states reorganized along linguistic lines. Rather, various combinations of national,

23 Federal (In)Stability in India 129 ideological (most notably the Communists), regional, and caste-based parties exist in most states (Manor 1995, ). Recognition of the linguistic cleavage has not been destabilizing. Using Lipset and Rokkan s framework, this study contends that if the states of India had not been reorganized along linguistic lines, then the linguistic cleavage would have become the primary cleavage for the party system and for other political movements. Rather than destabilizing the federation, linguistic reorganization undermined the conditions for a single, primary cleavage of linguistic politics to arise. The creation of homogeneous states provided the conditions for other identities such as region, caste, and class to come to the fore (Horowitz 1985, 617). 24 One of the reasons the Indian state has been successful in accommodating its diversity and maintaining its democracy 25 is because of these cross cutting cleavages (Manor 1996, 464). From this perspective, the dangers of having one cleavage in Indian politics become all too apparent. The promotion of multiple cleavages has a positive effect on federal stability because it provided a structural and institutional incentive for the proliferation of regionally based parties. Weiner saw this as a source of instability, arguing that the multiplicity of political parties threatens to destroy stable government (1957, 289), but stable government is not the same thing as a stable federation. 26 Multiplication of parties at the state level demonstrates that intra-elite competition exists. However, a proliferation of parties in a state will not contribute to federal stability if these parties are ethnically defined and seek to outflank each other (Nordlinger 1972, 118). In such situations, a proliferation of parties within the unit is likely to lead to conflict with the center. But this only happens when ethnically defined regional parties attempt to prove to their electorate that they can stand up to, or gain concessions from, the center. This situation is likely to occur when the identity around which the unit is defined feels threatened. In India, in contrast, the multiplication of parties has generally been indicative of federal stability because political parties have tended to not base their platforms on mobilizing against the center. Nehru and the Congress had to concede linguistic reorganization primarily for electoral reasons. The effective number of linguistic groups at the all-india level was too high to have done otherwise. The success of the reorganization, in terms of protecting Congress dominance, is illustrated by the fact that the numbers of parties dropped dramatically between the general elections of 1951 and The reduction in parties standing in the 1957 elections was partially a result of the perceived futility of standing against the Congress. It was also because the Congress temporarily reaped the benefits of reorganization and forestalled further factionalism. As Weiner noted at the time,

24 130 Federalism and Ethnic Conflict Regulation [a]gitation for the redistribution of states along linguistic lines has sharpened the factional divisions in some areas, but at the moment no other issue of state policy seems likely to provide a further basis for the further development of factions (1957, 283). In actuality, the Congress suffered a major split a decade later. If linguistic reorganization had not been conceded, then electoral considerations would have caused many more defections before this date. The demand for linguistic reorganization was an issue to create a coalition of voters that spans electoral constituencies (Chhibber 1999, 18). Nehru had sought to prevent linguistic identities from gaining representation in decision-making institutions, but as Kothari points out, soon after the successful culmination of the agitation the Congress absorbed a large number of the new entrants [into politics] and succeeded in capturing full initiative in state politics (1964, 1168). The primary cleavage was thus diffused. Table 6.8 demonstrates that there is a correlation between the number of political parties in the state and whether it is a Hindi-speaking state. Table 6.8 Indian general elections Correlation between enling and nenseats sorted according to the majority language spoken in the state Language Correlation Coefficient Sig. Sig. (1-tailed) N Non-Hindi-majority state ** Hindi-majority state ** Correlation is significant at the.01 level (1-tailed) Source: Electoral data adapted from David Butler et al. (1995), Mahendra Rana (1998), and Election Commission of India ( ). Linguistic data adapted from Government of India Census (1951, 1961, 1976, 1991). Notes: These data exclude the units of the federation that return only one seat to the Lok Sabha. These units skew the relationship as nenseats cannot be greater than 1.00, regardless of the score for enling. I use the log of enling and nenseats to enable me to use Pearson s Correlation. These data cannot prove that linguistic reorganization caused the increase in the effective number of legislative parties, but they are supportive of the argument. Multiple explanations for the post-1967 decline of the Congress have been advanced, all of which explain only part of a complex story. Kohli has discussed Indira Gandhi s role in undermining the organizational roots of the party (1991, 5 6); Brass has focused on the entrance of new social groups into the political system (1997, 204); and E. Sridharan focuses on the interactive nature of the federal and electoral systems. His explanation is concerned with the systemic properties of the first past the post electoral system working themselves out in a federal polity (Sridharan 2002, 495).

25 Federal (In)Stability in India 131 The institutionalist explanation at the heart of this study is premised on the importance of the type of federal system created in India. Linguistically homogeneous units have contributed to the proliferation of parties. This was the opposite of what Eric Nordlinger s thesis would predict. The fact that differences emerged between the Hindi-majority and non-hindi-majority speaking states was extremely significant. In a federal system those states not part of a heartland would be more likely to develop in opposition to the center. Instead, a multiplicity of parties emerged, and more significantly, multiple parties not seeking to ethnically outbid one another. These non-hindi-majority states were not part of the core and fought the attempt to introduce Hindi as the sole official language. During the era of Congress dominance all [Congress] Prime Ministers contested from constituencies in Uttar Pradesh at some time (Butler et al. 1995, 68). 27 Sonia Gandhi, the current president of the Congress was also elected from UP. 28 These non-hindi-speaking states initially fought the attempt to introduce Hindi as the sole official language. The fact that it was the non-hindispeaking states that initially divided into competing factions illustrates that linguistic reorganization proved to be an excellent accommodative strategy. In Madras there were profound tensions before reorganization; the fasting to death of Potti Sriramulu in order to secure a Telugu-speaking state in 1953 was the catalyst for the wider reorganization of states. The tensions that would have been caused by not conceding linguistic reorganization can be illustrated through a discussion of the Bombay state, excluded from the initial reorganization. In Bombay, after the denial of the claims for a bifurcation of the state, the Congress was challenged by two parties: the Maharashtra Samyuka Samiti (MSS) 29 and the Mahagujarat Janata Parishad. Michael Brecher notes that [l]eading Congressmen joined the chorus of dissent (1959, 484), despite the Congress Working Committee (CWC) calling on Congressmen to avoid the agitational approach (Windmiller 1956, 132). In November ,000 workers left their jobs in response to the call of the leftist leaders. Almost immediately, violence exploded over the city and an orgy of rioting and destruction resulted (Windmiller 1956, 135). Agitation on the Gujarati side was more muted, as a linguistic division meant losing the cosmopolitan business city of Bombay to Maharashtra, but [d]ay by day the situation became ugly opening out opportunities for people with fissiparous political views to create confusion (Roy 1962, 208). In January 1956 riots [had] enveloped Bombay [and] eighty people were killed and 450 wounded [in] the large-scale police firing (Brecher 1959, 484). 30 Although Nehru argued that the coming elections could go to hell. I am tired of listening to talks of pleasing this party and that party (Windmiller

26 132 Federalism and Ethnic Conflict Regulation 1956, 141), the state Congress government found it impossible to turn a deaf ear to this demand and carry on with the administration of the state insensible to the constant agitation (Roy 1962, 207). The MSS was very successful in the 1957 elections, demonstrating that sentiment for the creation of a Maharashtrian state is still strong. It reduced the Congress majority in the Bombay state legislature (Weiner 1957, 267). The effective number of legislative parties in Bombay in 1957 was the second highest in the country. This must be compared to other states where the Congress had retrenched its position after linguistic reorganization. Although there were no calls for the secession of the state, the violence was widespread. The instability led to the bisection of the state. The Congress reaped the benefit of reorganization in the 1962 election and won an easy victory over all the opposition parties (Joshi 1968, 194). Linguistic reorganization thus produced a similar result in this state to what it did in the rest of India five years previously. However, correlation cannot prove causation; federal design cannot account for the initial proliferation of parties in the non-hindi-majority states. As Wilkinson discusses, in the early twentieth century, the status of the lower castes in the southern, non-hindi-speaking area of India was much worse than in the north. In these southern states there had been caste agitation in favor of job reservations. The British were keen to concede these reservations because the Brahmins in these states were at the forefront of the Home Rule movement. As Wilkinson explains, this meant that the party systems in the southern states were initially more fractionalized than those in the north [b]ecause the colonial state provided institutional incentives for backward-caste mobilization, substantial intra-hindu party political competition emerged as early as the 1920s and 1930s (2004,173 74). Linguistic reorganization was therefore not the sole factor that caused the party system to fractionalize. The state of Madras would be an obvious counter example, as the Dravida Munnetra Kazagham (DMK) under C. N. Annadurai split from the Dravida Kazhagam in However, Tamil Nadu s (as it became) party system ultimately supports the hypothesis. Despite tensions with the center, the state strongly supported the Indian Union in its 1962 war with China. The DMK called off its secessionist campaign in 1963 on the grounds that there is a threat to our sovereignty and we must act as one. [O]nly if there is a state can we ask what we want of it, (Rajagopalan 2001, 156). While Manor questions the extent to which politicians and parties in Tamil Nadu were actually secessionist (2001, 87), the abandonment of the secessionist rhetoric and the split in the DMK in 1972 was significant. For a state with such an antipathy to Hindi speakers

27 Federal (In)Stability in India 133 and elements of Hinduism, 31 this was a definite sign of security within the federation. More importantly, in recent years Tamil parties have been prominent members of the coalition of parties at the center. If a Tamil state had not been created, and the debate over Hindi as a national language not reconciled to their satisfaction, a very different outcome would have been expected. Chart 6.2 Variance in the effective number of legislative parties (nenseats) in Indian general elections, Source: As Table 6.8. Chart 6.2 demonstrates that in recent years the differences in the number of effective legislative parties in the Hindi-majority and non-hindi-majority states have grown smaller. Northern Hindi speaking states have seen a rise in demands for job reservations for the lower castes. These demands provided the basis for the formation of political parties, such as the Bahujan Samaj Party (BSP). In 2004 UP was a battleground for 3.6 effective legislative parties, including the BSP, Samajwadi Party (SP), and the BJP. Linguistic reorganization was not the sole cause of fractionalization; it facilitated other cleavages mobilization. As Ashutosh Varshney discusses, these lower caste communities have articulated a very different version of the Indian nation than either Hindu or secular nationalists.

28 134 Federalism and Ethnic Conflict Regulation Rather than talk about the nation and the placement of religious or linguistic groups therein, the caste narrative speaks of the deeply hierarchical and unjust nature of Hindu social order. [A]n egalitarian restructuring of the Hindu social order is the chief goal [his emphasis] (2002, 57). The Hindi-speaking states were already relatively homogeneous along linguistic lines. They did not factionalize immediately, as Weiner observed, because the population identified with all-india parties more than those states that did not have a Hindi majority. In the Hindi-speaking areas there tends to be less of an identification with the State governments as such, compared to the degree of identification among the non-hindi language many Hindi speaking politicians see the Hindi region as the heartland of India (1967, 325, 342). As Hindi speakers are the largest linguistic group within India, these states can be understood as members of the staatsvolk. 32 The staatsvolk members of a state typically do not perceive any incompatibility between the identity of the region and the identity of the state for example, the English in Britain and Punjabis in Pakistan. The rise of political parties such as the BSP and SP is indicative of the fact that they do not identify with the staatsvolk to the same extent, which explains why this process no longer operates. Conclusion The argument above is necessarily incomplete. An analysis of the number and types of political parties as a proxy for federal stability cannot take into account the numbers of movements that articulate their demands outside constitutional structures, or that boycott specific elections because of their concerns about the legitimacy of the process (to say nothing of their concerns regarding the legitimacy of the state). Although India, unlike Pakistan, has not experienced a successful secession it has deployed extensive force to maintain its territorial integrity, questioning the reality of its stability. India has been more conflict prone than Pakistan (Gurr 1993, ). Yet India s population is much larger than that of Pakistan, and much more religiously and linguistically diverse. Using this measure on its own would ignore the real success that India has had accommodating linguistic identities, though it has not been as successful at accommodating religious ones. This is demonstrated in Chart 6.3. As Chart 6.3 reveals, incidents that were mobilized around linguistic issues have declined since the 1950s. Linguistically homogeneous units have

29 Federal (In)Stability in India 135 created conditions for alternative identities to thrive. That linguistic identities remain strong despite linguistic reorganization is not disputed; one only has to look at Tamil Web sites to see a vociferous campaign against insipid Hindi-ization. 33 Yet linguistic identities have generally become compatible with, rather than antithetical to, Indian identity. Federal structures were not sufficient to promote this security; linguistic accommodation at the center was also essential. Chart 6.3 Sample of all incidents of ethnic mobilization in India, Source: Wilkinson (2002, 15). Notes: The cases are taken from a sample of days in each year. The absolute numbers cited are therefore lower than the actual number, but their relative position is accurate. These data are very different in relation to religious mobilization. Mobilization around a Muslim identity was nonexistent in the 1950s. Since its emergence in the 1960s, it has increased twelvefold and has been especially important in the long-running insurgency in Kashmir (Wilkinson 2002a, 15). 34 Incidents that were mobilized around Sikh identity increased twelvefold from the 1950s compared to the 1980s, dropping back slightly in the period. Federal homogeneity cannot account for the conflict in the non-hindu-majority states, as these states are not homogeneous along linguistic or religious lines. Indeed, that has been one of the factors promoting insecurity. The policies of the center have been vitally important in understanding the tensions that arose in the states of Punjab, Kashmir, and the Northeast all of whose conflicts were escalated by central policies and interventions. As Subramanian notes, [t]he creation of a state for Punjabi speakers and the added autonomy given to Jammu and Kashmir failed to satisfy ethnic mobilizers, partly because the